Below is a drone photo I took that was in very low light. The more I used it, the more I could see how it could easily fit into my workflow. Some of this can be done more traditionally with gradient masks, but I like the flexibility this relight feature offers. The feature also includes the ability to warm up parts of the image, and using depth cues from the image, you can warm up the foreground, background, or both. That allows you to be very selective about what parts of the image get highlighted. With Luminar Neo, the relight commands let me bring up the level of the foreground and lower the level of the sky. Normally, I would bring the highlights down to preserve the skies and open up the shadows. I opened some drone images taken in low light. Of course, it makes a mistake, you're just a click away from a selective erase and restore tool that can fix any AI missteps manually. In the Skylum demo we saw online, it left faraway birds in the sky while removing sensor dust. One of my images had an airplane flying over, and Luminar Neo knocked out the dust but left the airplane alone. I felt the AI software was very smart about what to remove and what not to remove. I keep my sensors pretty clean, but I did find some old raw files that had some sensor dust. For my testing, I worked on a brand new MacBook Pro with 32 GB of RAM and an Apple M1 Max chip. We were told to expect a few glitches, and the speed would not be representative of its performance on Apple or PC hardware. I want to preface my comments with a caveat.
I could open images, let the new features do their work, and then export the edited file as a 16-bit TIFF, then continue editing in Photoshop or the current Luminar AI.
The software we received for early testing included just those features, so it was not a complete editing solution.
The features I tested include a relight feature, sensor dust removal, and power line removal. This week, Skylum got some reviewers together online, myself included, to get a briefing on three particular new AI features in the upcoming product, and then turned us loose with an early beta so we could see how it all worked.
0 kommentar(er)
